Categorizing Unemployment and Policy Responses

About a week ago, I wrote in response to a remarkably venal column by an economics professor the value of seeking Pareto improvements in deals and paying off the losers from the additional surplus society has gained:

the cost of unemployment due to structural outsourcing from the combination of direct wage losses, direct re-training/job seeking expenses and indirect future losses due to the combination of a lower base wage and lower future trend growth

The Washington Post had an interesting article this weekend concerning long term unemployment and how even in a 'good' labor market that we had until recently as measured by the unemployment rate, structural unemployment has been increasing:

An unusually large share of workers have been out a job for more than six months even as overall unemployment has remained low....In November, nearly 1.4 million people -- almost one in five of those unemployed -- had been jobless for at least 27 weeks, the juncture when unemployment insurance benefits end for most recipients. That is about twice the level of long-term unemployment before the 2001 recession....

Once concentrated among manufacturing workers and those with little work history, education or skills, long-term unemployment is growing most rapidly among white-collar and college-educated workers with long work experience, studies have found, making the problem difficult for policymakers to address even as it grows more urgent.

The growth in long-term unemployment has occurred even as displaced workers have taken bigger pay cuts to reenter the job market. A 2004 study found that workers who lost a job in 2001 to 2003 took an average pay cut of 17 percent in their new jobs, more than double the average cut of those displaced in the late 1990s

So what are the appropriate policy responses to unemployment? What are the different types of unemployment that need to be considered? A gliberterian response is nothing is the appropriate policy response and the markets will clear once an individual's reservation price drops low enough, and if that price is below minimum wage, we should abolish minimum wage laws. Now that I am done beating that strawman to death, let's look at different types of unemployment.

The first is incidental/voluntary unemployment when someone leaves a job on their own free will and with no co-ercion on the part of the employer, management or co-workers. For instance when I was a dishwasher at a local bar, I left after I received a student loan windfall that was sufficient to cover my expenses for the rest of the year. The correct policy resopnse is to do very little here in my opinion unless there is clear evidence of co-ercion, discrimination or harassment, and then legal channels for dispute resolution and restitution are the proper channels.

Frictional unemployment will exist as people shift between jobs, people enter and leave the labor force, and people wait for work to get started. If there is high certainty that work will be starting in three weeks, the cost of finding and starting a temporary job for two weeks tends to be high enough to deter that course of action.

The next type of unemployment is cyclical unemployment. This is predictable unemployment where jobs temporarily go away in the field where an individual has a high degree of investment, skill, training and human capital. The classic example is the two week shutdown at a large steel mill during the summer for preventive and recurring maintenance. The plant sees 90% of its work force idle but everyone knows that in two weeks the plant will be back at full capacity. In this short term cyclical unemployment, minimal intervention is usually required as everyone knew about it in advance, and were able to plan around it. Frequently this planning includes taking voluntary or mandatory paid vacation or other compensation shifting. Another cyclical example is when companies cut back as revenue comes in under estimates. The job may disappear at that particular company but the job is still somewhere within the regional/national economy and there is just a matching problem. The public policy response is often very limited unemployment benefits as most states have at least a one week elimination period before anyone qualifies for benefits.

Policy becomes much more involved at longer term cyclical unemployment. My dad has worked in the construction industry for almost his entire adult life, and this is something that we have experienced several times as a family. The industry as a whole is down for a while, and there is no hiring, and no big projects are in the immediate pipeline. However, in the long run, there is work available and projects will start (eventually it seems at times). Bridging the gap between nothing now and plenty of work in the future takes the form of extended unemployment benefits, training for other opportunities to thin out the sector specific labor market. The goal here is to ameliorate pain as well as even out the labor pool so that there are sufficient people available for the boom times so that wage inflation does not take off.

Sometimes long run cyclical unemployment turns into structural unemployment where an individual's skill set and human capital do not map onto any locally available work at anything vaguely approximating their previous wage and permanent income assumptions. This is where significant policy intervention, in my mind, is justified. Retraining assistance, health-care subsidies and bridge programs as well as cash assistance will be appropriate. Wage insurance as a bridging program would be a good idea as people would have the resources to look for a new career. Society as a whole has a massive risk leveling ability that any individual does not have, especially if we assume that structural unemployment is a non-homogenous event and concentrates in specific localities at specific points in time. If we assume this, then local communities and individuals within these communities will be hit with systemic shocks that personal and local reserves will have a very difficult time covering.

Structural unemployment is very expensive for individuals and regions to bear. For instance Pittsburgh looks like it may soon be bottoming out from the structural loss of the integrated steel industry and this region has taken on massive per-capita debts and seen mediocre generational wage and job growth. A steel worker laid off in the massive waves of the early to mid-80s saw one of his main assets, his house, depreciate at the time that he needed access to his accumulated savings to change jobs, change careers and probably change locations.

There are other sorts of unemployment, and plenty of other means of policy intervention including the use of the disability insurance system as a means of reducing the labor force during tough times, as well as increased early retirements/buy-outs. And then there is the gray and black markets as unemployment responses such as working under the table, cash only to sweatshops manned by immigrants. But these basic types of unemployment cover most situations and describe the basic policy arrangements.